Corruption is one of the several unwanted constants around the world every year. According to an annual report created by Transparency International, in 2020, corruption was still widely prevalent—and, in several nations, rampant—across the world. What’s worse, the study also found that corruption affected and, in many cases, slowed down the global response to Covid-19. Unfortunately, the point about corruption and catastrophes feeding off each other still holds true to this day. While corruption in certain countries may be higher than in others, it is still everybody’s problem to solve. Eliminating it completely is impossible as it is too ingrained in our social fabric now.
Modern technology provides some hope in our collective crusade against global corruption. For example, the usage of AI in smart cities shows how governments can disperse citizen services efficiently with minimal to non-existent corruption present in the process. The potential of AI to detect and, eventually, reduce corruption in various places of power is high. Therefore, a problem like corruption needs to be handled with AI, along with other elements to support it.
Data Mining to Combat Corruption
Unsurprisingly, if AI is used as an anti-corruption tool, big data and analytics will contribute heavily towards the process too. Here’s a summary of how big data can assist AI with combating corruption: Initially, big data will play a significant role in collecting and compiling corruption-related information from various sources. This information is then fed into AI systems for processing and predictive analysis.
Mining is the process of collecting data in massive quantities from relevant sources before an AI system looks for patterns and anomalies within the data to find specific relationships between variables.
The use of data mining in the ongoing data science revolution can be likened to the use of coal mining playing a key role throughout the Industrial Revolution. But how does data mining resolve the corruption problem?
In the last decade, Transparency International Georgia, a prominent anti-corruption body, launched an open-source analytics and procurement monitoring web portal that allows users to get information regarding the Georgian government procurement processes from official sources. Using that information, Georgian citizens can check specific procurement transactions, the names of the companies bidding for public contracts, and other statistical and factual data regarding how their government spends the taxpayers’ money. Such easy access to data of that importance empowers curious individuals in the country to check for suspicious activity within the electronic tender-related information. If legal violations or possible corruption by governments are detected, people can submit an online complaint that will be seen by a specialized dispute resolution board. The board then goes on to review and address such complaints within ten days. Data mining is the driving force behind the collection of government procurement information. The incoherent big data generated from mining is used by AI, which understands, simplifies and structures information into user-friendly modules for the regular folk to read and comprehend.
Data mining can be used to spot inconsistencies in the various recorded operations involving elected government authorities. The inconsistencies found are related to how their government spends public money.
Once the mining phase is complete, data analytics takes over. Analytics is used to detect fraud or other irregularities in private and public bodies’ online documents and websites. To fight corruption in procurement and other government-related activities, it is essential that accurate, verified information is uploaded onto official government websites consistently. Third-party auditors and data specialists must be hired to carry out the verification of all information uploaded online.
Data mining and AI are useful for tax fraud detection, improvement of taxpayer compliance, money laundering and other types of fraud committed by politicians and other officials connected to the political spectrum.
As we can see, data mining and AI form a potent combination to facilitate the detection and prevention of corruption-related activities in the field of public procurement and a few others.
AI and Blockchain to Reduce Corruption in Pharmaceuticals
Blockchain and AI share an interesting dynamic. As we know, both technologies have their own share of weaknesses. While blockchain systems demonstrate issues related to data scalability and efficiency, AI has shortcomings in areas such as explainability, data privacy and trustworthiness. When the two technologies are used together, they cover each other’s deficiencies to form a potent combination. Blockchain brings explainability and privacy in AI systems, while the latter empowers blockchain systems to achieve greater scalability that can be useful for performance boosting, governance and personalization.
Corruption is not just related to the political landscape or, in general, only the public sector. It can be found in the private sector too. In the last decade, the pharmaceutical industry had to deal with corruption allegations, and it continues to do so. When instances of corruption and fraud are found in big pharma—or any other—major private corporations, there are massive penalties, fines and other repercussions levied at the entity responsible for their possibly corrupt actions. The pharmaceutical industry is often accused of valuing profits more than the lives of actual patients in a hospital.
AI and blockchain systems can help to address issues of corruption and fraud whenever they come up in pharmaceutical companies. Additionally, the information available to patients regarding health conditions and medicines can be improved with AI and blockchain systems in the healthcare industry. Patient knowledge is vital to prevent a specific type of corruption in the pharma industry. Pharmaceutical companies may bribe reputed physicians so that they can push their products on unsuspecting customers. So, customers looking for a specific medicinal product may end up purchasing the product recommended to them by the physician. As patient knowledge improves with clearer information, they do not have to depend on physicians who may try to push certain medicines and products just so that the pharma company making them can increase their own revenue.
In many ways, the corruption in pharmaceuticals and other private sector bodies may be different from our general idea of the concept of corruption. By using blockchain and AI, pharma companies are prevented from paying bribes to physicians simply to promote their products.
Blockchain and AI play their part in fostering faster information flow and data security so that such instances of corruption can be avoided by pharma companies. The increasing presence of AI in smart cities can reduce corruption in the private sector as well as the public sector.
Predictive Analysis for Forecasting Corruption
A few researchers from the University of Valladolid, Spain, have developed a concept AI model with artificial neural networks that can forecast corruption cases that could take place after one to three years. For example, the data generated from the system can check factors such as real estate tax in an area, a dramatic rise in housing prices there, the creation of brand-new companies in the area and the opening of bank branches there to predict a probable public corruption-related activity for the region.
To do the study and create the system, the researchers used datasets that included every case of corruption that took place in Spain between the years 2000 and 2012. Predicting corruption can be truly challenging for AI systems, and the usage of predictive analysis—using past data to forecast future occurrences—is a decent idea if systems like these are used extensively in the future.
The advancements in technology and the presence of AI in smart cities can allow systems like these to flourish.
Challenges In Implementing AI to Fight Corruption
As ever, any technology application in a conceptual stage will come with its own set of issues. For AI, the two main challenges in its implementation as an anti-corruption tool are:
a) A lack of explainability
b) A lack of quality datasets
The lack of explainability will leave network administrators and AI experts in hot water as they would not be able to justify certain AI decisions. So, using a combination of AI and blockchain helps, as stated above.
The problem of dataset shortage can be resolved if the experts training such systems can get their research right and attain information regarding most—if not all—corruption cases in the recent and more distant past and include them to beef up the training datasets.
The use of AI in smart cities and elsewhere can be useful to deal with corruption. However, as we have seen, many elements have to come together for the implementation to work successfully, and the process of implementing AI to curb corruption needs to be properly thought through by researchers and others involved.